code-semantic-search

Pass

Audited by Gen Agent Trust Hub on May 10, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill describes a RAG-based search system (Indirect Prompt Injection) which ingests untrusted data from source-code corpora.
  • Ingestion points: Processes code files from the local repository as described in the overview and instructions.
  • Boundary markers: The instructions do not include the use of delimiters or 'ignore embedded instructions' warnings for the retrieved content.
  • Capability inventory: The skill uses tools like Bash, Read, Glob, and Grep which can be leveraged if malicious instructions are encountered in the processed code.
  • Sanitization: There is no mention of sanitizing or escaping the retrieved code snippets before they are presented to the agent's context.
Audit Metadata
Risk Level
SAFE
Analyzed
May 10, 2026, 05:22 PM
Security Audit — agent-trust-hub — code-semantic-search